The present invention generally relates to electronic communications and, in particular, to devices that generate digital signatures associated with electronic communications.
As used herein, an electronic communication (“EC”) is considered to be any communication in electronic form. ECs have become an integral part of transacting business today, especially with the growth of the Internet and e-commerce. An EC can represent, for example, a request for access to information or a physical area, a financial transaction, such as an instruction to a bank to transfer funds, or a legal action, such as the delivery of an executed contract.
Over recent years, digital signatures also have become an important part of e-commerce. The origination of a digital signature generally comprises: (1) the calculation of a message digest-such as a hash value; and (2) the subsequent encryption of the message digest. The message digest is encrypted by an electronic device generally using a private key of a key pair used in public-private key cryptography (also known as asymmetric cryptography). The resulting ciphertext itself usually constitutes the digital signature, which typically is appended to the message to form the EC. The second part of originating the digital signature-encrypting with a private key-is referred to herein as “generating” the digital signature, and the combination of the two steps is referred to herein as “originating” the digital signature. Furthermore, while the generation of the digital signature is conventionally understood as the encryption of the message digest, it is contemplated herein that generating the digital signature also may include simply encrypting the message rather than the message digest. Digital signatures are important because any change whatsoever to the message in an EC is detectable from an analysis of the message and the digital signature. In this regard, the digital signature is used to “authenticate” a message contained within the EC (herein referred to as “Message Authentication”).
For example, a message digest may be calculated by applying a hashing algorithm to the message. The hashing algorithm may be applied either within the device or external to the device with the resulting hash value then being transmitted to the device for generation of the digital signature. In order to perform the Message Authentication in this example, the recipient of the EC must know or be able to obtain both the identity of the hashing algorithm applied to the message as well as the public key (“PuK”) corresponding to the private key used to encrypt the message digest. With this knowledge, the recipient applies the appropriate hashing algorithm to the message to calculate a hash value, and the recipient decrypts the digital signature using the public key. If the hash value calculated by the recipient equals the hash value of the decrypted digital signature, then the recipient determines that the content of the message contained in the EC was not altered in transmission, which necessarily would have changed the hash value.
In performing Message Authentication, the recipient also authenticates the sender of the EC, in so much as the recipient thereby confirms that the sender of the EC possessed the private key corresponding to the public key used successfully to authenticate the message. This is one type of entity authentication and is based on what the sender “has” (herein referred to as “Factor A Entity Authentication”). Factor A Entity Authentication is useful when the recipient of the EC has trusted information regarding the identity of the owner of the private key. Such trusted information may arise from a digital certificate issued by a trusted third-party that accompanies the EC and binds the identity of the private key owner with the public key. This trusted knowledge also may comprise actual knowledge of the identity of the private key owner, such as in the case where the recipient itself has issued the private key or the device containing the private key to the owner.
As will be appreciated, trust in the digital signature system depends upon the legitimate possession and use of the private key, i.e., upon the sender of the EC actually being the private key owner. A fraudulent use of a private key to generate a digital signature of an EC currently cannot be detected through the above-described Message Authentication and Factor A Entity Authentication procedures. The digital signature system therefore is susceptible to fraud if a private key of a device is stolen, either by physical theft of the device containing the private key, or by discovery of the private key therein and subsequent copying and use in another device capable of generating digital signatures.
To guard against fraudulent use of a device through theft of the device itself, a personal identification number (PIN), password, or passphrase (collectively referred to herein as “Secret”) is typically prestored within the device and must be input into the device before it will operate to generate digital signatures. Alternatively, the Secret is shared with the recipient beforehand and, when the EC later is sent to the recipient, the Secret also is sent to the recipient in association with the message. In the first case, verification of the Secret authenticates the user of the device (herein “User Authentication”), and in the second case, verification of the Secret authenticates the sender of the EC (herein “Sender Authentication”). In either case, confirmation of the Secret represents entity authentication based on what the user or sender “knows” (herein “Factor B Entity Authentication”). Another security feature that guards against fraudulent use of the device through physical theft include the verification of a biometric characteristic-like a fingerprint-of the user of the device or sender of the EC. This type of authentication is based on what the user or sender “is” (herein “Factor C Entity Authentication”). As with the Secret, a biometric value is either maintained within the device for User Authentication, or is shared with the recipient beforehand for Sender Authentication by the recipient. To guard against discovery of a private key and subsequent copying and use in another device, devices are manufactured with electronic shielding, zeroization, auditing, tamper evidence and tamper response, and other security features that serve to safeguard the private key (and other protected data) contained therein.
Such security features of devices include hardware, software, and firmware, and are well known in the art of manufacturing secure computer chips and other devices having cryptographic modules. The requirements of such security features are specified, for example, in Federal Information Processing Standards Publication 140-1, Security Requirements for Cryptographic Modules, US DOC/NBS, Jan. 11, 1994 (herein “FIPS PUB 140-1”), which is incorporated herein by reference and which is available for download at http://csrc.nist.gov/publications/fips; and Federal Information Processing Standards Publication 140-2, Security Requirements for Cryptographic Modules, US DOC/NBS, May 25, 2001 (herein “FIPS PUB 140-2”), which is incorporated herein by reference and which is available for download at http://csrc.nist.gov/publications/fips. FIPS PUB 140-1 and 140-2 also define security levels that may be met by a device based on the device's security features, with each of these defined security levels generally representing a various level of difficulty-in terms of time and money-that would be encountered in attempting to discern a private key of a device. Currently, four security levels are defined with security level 4 being the highest level of security available.
Specifications for such security features also are set forth in Trusted Computing Platform Alliance Trusted Platform Module Protection Profile Version 0.45, TRUSTED COMPUTING PLATFORM ALLIANCE, September 2000; Trusted Platform Module (TPM) Security Policy Version 0.45, TRUSTED COMPUTING PLATFORM ALLIANCE, October 2000; and TCPA PC Implementations Specification Version 0.95, TRUSTED COMPUTING PLATFORM ALLIANCE, Jul. 4, 2001, which are incorporated herein by reference (collectively “TCPA Documents”), and which are available for download at http://www.trustedpc.com; and Common Criteria for Information Technology Security Evaluation, Smart Card Protection Profile, Draft Version 2.1d, SMART CARD SECURITY USER GROUP, Mar. 21, 2001, which is incorporated herein by reference (hereinafter “Smart Card Protection Profile”), and which is available for download at http://csrc.nist.gov.
The particular features of a device that safeguard against discovery of a private key and other protected data are referred to herein as “security characteristics” of the device. The particular features of a device that safeguard against unauthorized use of the device by authenticating the user are referred to herein as “authentication capabilities” of the device. The “security features” of a device (including a cryptographic module or TPM) comprise the security characteristics and authentication capabilities as well as other security features of the device, the requirements of which are specified in the above cited references.
Unfortunately, while the aforementioned security features generally reduce the risk of fraud within the digital signature system overall, a recipient of any one particular EC including a digital signature may be unfamiliar with the device used to generate the digital signature and, therefore, be unable to gauge the risk of whether the digital signature was generated fraudulently, either through theft of the device or discovery of the private key.
Of course, if the recipient possesses a shared secrete or a biometric value of the sender for performing Sender Authentication, then the recipient may determine that the digital signature was not fraudulently generated assuming that the shared secret or biometric value was not stolen. However, this type of protection by the recipient has significant drawbacks and is not always used by the recipient. For example, if the Secret or biometric value is communicated to the recipient in association with a message for Sender Authentication by the recipient, then the Secret or biometric value first must have been shared with the recipient beforehand and safeguarded by the recipient as part of an established, preexisting relationship; consequently, a recipient having no prior existing relationship with the sender is unable to perform Sender Authentication.
Another drawback is that the sharing of the Secret or biometric value with the recipient exposes the recipient to liability and exposes the Secret or biometric value itself to a greater risk of theft and dissemination. The transmission of the Secret or biometric value for verification carries with it the risk of interception and discovery during transit. Upon receipt, the Secret or biometric value must be safeguarded by the recipient, which inherently gives rise to a risk of theft from the recipient. This is especially significant in the corporate context where a rogue employee may steal the safeguarded Secret or biometric value (insider fraud historically has been the greatest threat). The potential damages also are extensive when the Secret or biometric value is stolen. Since it is difficult for an individual to remember multiple Secrets for multiple recipients, it is common for the same Secret to be used with different recipients. The theft of the Secret from one recipient thereby compromises the Sender Authentication performed by all of the recipients, at least until the Secret is changed for each recipient. In the case of the theft of a biometric value, the damages are even more severe, as a sender's biometric characteristic cannot be changed and, once lost, potentially compromises any future Sender Authentication therewith.
Accordingly, a recipient generally is unable to gauge the risk of whether a digital signature was generated fraudulently when no secret or biometric value is shared between the sender and the recipient. Instead, a recipient must rely upon blind trust in accepting that the device used to generate the digital signature has not been stolen and in accepting that the device used to generate the digital signature has sufficient safeguards to protect its private key from discovery and use.
A need therefore exists for a method by which a recipient may reliably identify a risk of whether a digital signature has been generated fraudulently using a stolen private key (whether stolen by physical theft of the device or discovery of the private key itself), whereby the recipient may protect itself against fraud. In this regard, a need also exists for a method by which a recipient of an EC including a digital signature may reliably determine at any given time the current level of security of the device to which belongs the private key used to generate the digital signature. A need also exists for a method by which a recipient of an EC may reliably determine the safeguards of such device that protect against fraudulent use thereof.